Answers are grounded in this paper and cite their sources.
Genotypic tools based on the HIV-1 V3 loop, particularly Geno2pheno, are widely used to predict co-receptor tropism and guide maraviroc-based antiretroviral therapy. However, their predictive accuracy for non-B subtypes prevalent in China — especially CRF01_AE — remains uncertain. We therefore evaluated the performance of Geno2pheno against phenotypic validation methods in a cohort of primary HIV-1 clinical isolates from China. Thirty-seven primary HIV-1 isolates from patients at Ditan Hospital, Capital Medical University, were obtained by peripheral blood mononuclear cell (PBMC) co-culture. Viral RNA was extracted and V3 loop sequences of the env gene were amplified and sequenced. Geno2pheno was evaluated at both 10% and 5% false-positive rate (FPR) thresholds for genotypic tropism prediction. Phenotypic tropism was determined by GHOST-CCR5 and GHOST-CXCR4 cell infection assays, corroborated by MT-2 syncytium formation assays. V3 loop net charge, key amino acid residues, and N-linked glycosylation site polymorphisms were subsequently analysed to characterise the molecular basis of prediction discordance. Geno2pheno was evaluated at both 10% and 5% FPR thresholds. At 10% FPR, 11 strains (29.73%) were classified as CXCR4-tropic (X4) and 26 (70.27%) as CCR5-tropic (R5). Phenotypic assays identified 33 strains (89.19%) as R5-tropic, 3 (8.11%) as dual-tropic (R5 × 4), and 1 (2.70%) as strictly X4-tropic. Considering dual-tropic strains as concordant with X4 genotypic prediction because the genotype identifies the capacity to use CXCR4 regardless of concurrent CCR5 usage, the overall genotype–phenotype discordance rate was 18.92% (7/37). All 7 discordant strains were predicted as X4 by Geno2pheno yet phenotypically confirmed as R5-tropic; 5 of these (71.4%) belonged to the CRF01_AE subtype. Notably, at a 5% FPR threshold, all 7 discordant strains remained classified as X4 (FPR score range 0.2%-3.9%), indicating that the misclassification of CRF01_AE strains is not ameliorated by threshold tightening alone. The mean V3 loop net charge was significantly higher in the dual-tropic (5.33 ± 0.58) and discordant (5.29 ± 0.76) groups than in the concordant R5-tropic group (3.31 ± 1.12; p < 0.01). Discordant strains exhibited elevated V3 loop net charges (mean 5.29 ± 0.76) and heterogeneous substitutions at key positions relative to the HXB2 reference, including the retention of basic residues at position 11 in the B subtype strain DT105, yet functionally demonstrated CCR5 tropism. In this cohort, Geno2pheno tended to overestimate CXCR4 tropism prediction, with a prediction discordance rate of approximately 18.92% after accounting for dual-tropic strains as genotypically concordant, particularly in CRF01_AE strains. The MT-2 syncytium formation assay showed high concordance with GHOST cell validation, confirming its utility as a complementary functional method. For strains genotypically predicted as CXCR4-tropic — especially CRF01_AE — phenotypic validation is strongly recommended to ensure accurate tropism assignment and optimal clinical treatment decisions. Future efforts to develop and validate subtype-specific genotypic algorithms, including recalibration of Geno2pheno thresholds for CRF01_AE in larger, geographically diverse cohorts, are warranted.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: